4 Easy Ways to Launch Your First AI App TODAY

in this video I'm going to be covering the four easiest ways that you can create and deploy your very first AI powered app using gpt3 or other large language models I'm going to be taking four different methods that I've found and have been using recently that allow me to go from idea to prototype to putting it on a site as quickly as possible but also are not very Tech heavy and don't require you to know too much about developing and coding and in general hosting these kind of applications if you're someone with an idea but you just don't know how to bring it to Market or even create a prototype then this is going to be the video for you because I'm going to make this as easy as possible for people who don't necessarily have a technical background so that you can take your idea write your prompt host it and start playing around with it and see it come to life before your very eyes you're going to want to stick around to the end of the video because I'm going to be starting with the creation side of things and ending with the deployment so I'm going to start with showing you how you can create API endpoint and at the end I'm going to show you how you can connect the front end to their endpoint so you can start sharing it with other

people tool number one is everyprompter. com every prompt is like the improved version of the open AI playground this is where to start when you're building your first AI application now there's a number of features that the playground simply doesn't have and it's also a much nicer UI and easier to use so here you can see it's a typical playground environment with a model selector here a stop sequence enterer and temperature all your regular settings you can play around with in the playground then you have your prompt entry box here you have just put in a basic prompt of write me five YouTube titles for my video talking about the four best ways to create and deploy your first AI powered app and here it comes back with the response in this black section down here this is better than the playground for a number of reasons but this side panel here is one of them you can see that I can easily switch between different functions that I've created I can open a new one and then it will save them all for me to easily go back and forth between here is the stock Mentor prompt that I wrote from my previous video where I created an entire application using only no code tools so if you haven't checked that out I'll put a link here it's really good video you should definitely check it out you may have seen in that video one of the things about every prompt is that you can insert Dynamic values into your prompts which is essentially earmarking it for later using it in your application now you can see here if I enter this double curly brackets with input in between then on the right we get an input box here now anything I enter in the input box when I submit it it's going to inject it in I miss my family an injector in there and then we get the response we have two ears in one mouth so that we can listen twice as much as we speak so the best thing about every prompt is that now that I've completed my prompt that I'm happy with that I have my little Dynamic path put in I can literally one click deploy this so that I can start interacting with this uh programmatically through a website up here on the right you can see deploy and here it is active deployment I can view the API documentation and what this has done is packaged it all up and I can change the language here to python so that we can have a play around with it now so what this has done is taken this entire font that I wrote and then combined it with their backend system for handling this and created an API endpoint for me to contact with the post request and then also here you can see the input value is the key that I put in the dynamic and the double curly braces so all I need to do when I'm interacting with this is change this example value and then send it off to this API endpoint it also shows us the expected response so here I can very clearly see that I'm going to be wanting to go on to the choices array and then pluck out this text value so this looks a bit Cody I know but I'll show you in a second what it looks like when you're actually running it now I've hopped over to collab just to pop this into a notebook and show you guys how this works so here is the input value I've changed it to my dog died and I miss him a lot so if I submit this it's going to connect to that API endpoint send off the information and get the response back and here you go we have the choices now this looks a bit complicated but what I can do is pluck those out and get the actual text value that we want and just like that we have a custom API endpoint that we can send off our user input and get back a customized response from the API the response to my dog died and I missed him a lot is the key is to keep company only with people who uplift you whose presence calls forth your best from Epictetus now that we have a way to interact with our prompt insert the user input what we need to do is connect this to a front-end so that we can start to interact with it like a normal website or an application so I'm going to be covering all of that at the end of the video so stay tuned so the next way that

you can create and deploy your very first AI application is with a tool called promptable. ai now at a first glance this may look very similar to every prompt but this is actually a little bit different and I want to show you why it's similar to every prompt and that I can change between different prompts on the side here but what I want to do here is really show you this evaluate tab so here I have the stock mental quote from my previous video I've played around with it a little bit and I have my input value set here now if we head to the evaluate tab we can do some very important testing for our prompt it allows me to put in multiple different variables in place of that Dynamic variable of input within the prompt and allows me to essentially run them all at once and then see what the output is on mass I was blown away when I stumbled across this feature because this is such a crucial part of testing your prompts and of your AI applications so when you create some kind of application like the stock Mentor video I did previously and you seem to get the same response back quite a lot from the AI this kind of input and output testing is so valuable because I can start to see how often it's repeating the same thing if you're not familiar with this prompt I'll just quickly run you through it I'm telling it that it's a stoic life coach and it's gonna basically I'm a client and it's helping me with my life and it's going to give me a epictetious Marcus release CNN quotes back in response to the thing that I'm struggling with so essentially asking me what am I struggling with and I enter into that input field what I'm struggling with and then what it returns back is that Marcus Aurelius Epictetus or Seneca quote that is appropriate for that situation so I went over to chat gbt and asked it for 10 20 examples of things that people may be putting into this bot so I've got some here I'm feeling anxious and overwhelmed I'm not sure why having trouble dealing with that so I hope bunch of different potential things that people may be asking it so I've just cleared the output so you guys can see very clearly when everything comes in so I'm going to run this with the current prompt which I've included a whole bunch of quotes I'm going to run this prompt and see what it gives us now just at a glance I can already see that I'm getting the same quote repeated a few times so this having this kind of information about what this is spitting back on a large scale allows me to go back to my prompt and keep tweaking until I'm starting to get kind of responses that I want on Mouse this is why I love promptable because I've been able to come in here and create a really good prompt put in some Dynamic values that I can sort of earmark for using in my application head over to the evaluate Tab and start to see what some if we're getting similarities within the output and what kind of responses I get for a certain kind of input on mass saves me a ton of time and then I can head over to the deploy Tab and it literally deploy my application in a clip all I need to do is hit add deployment we'll call it video once again we have an endpoint that we can interact with a get request so we're going to take this and put it into our front-end application and I'm going to be showing you at the end of the video how you can interact with these uh apis

essentially that you're creating the third way that you can create and deploy your very first AI application is with steamship is essentially a managed backend for AI services so that you can come in and spin up a quick application and then they'll create and host it for you so that you can interact it with it programmatically from your front end this is more versatile you get more complex than promptable in every prompt because you can essentially run a prompt based apis but you can also build Lang chain apps which is sort of the major Focus point for people building on large language models at the moment and I'm going to be making a video on it for you guys very soon so stay tuned for that you can either run these from your machine or from replit I'm going to use replit and what we can do is clone their Ripple they give you an entire template that you can use you simply have to clone it and don't freak out because this looks like a lot of code if you're not very technically inclined this is actually very basic stuff nothing too fancy here and if you really wanted some help with whipping together something like this you can find any python developer on upwork for ten dollars an hour and he'll be able to help you out and get this done I've just quickly spun up a quick example of an application you can make with steamship now this is just a really gimmicky stupid thing to do but I thought I'd just make you guys an example so what this app does is take in your name and your age in response that you get back from is essentially some advice based on your age so what I've asked her in the prompt is do you have any advice for insert age you're old like me so it's going to give you back advice based on your age it goes through here nothing too fancy if I run this you can see on the side here generate personalized advice for gpt3 and I need to go into my API key let's run through an example it's taken in John his age is 35 and the advice of the AI is given back is my advice would be to continue learning and growing stay curious and find ways to charge yourself so if I go Liam 22. my advice for a 22 year old like you is to make the most of your 20s great the amazing thing about steamship is that once you've written your application you've got the functionality working all I've needed to do is run the ship deploy command and then it gives me this right here which allows me to interact with that endpoint that they've created within their website now this is just an example on their website does the same functionality that I just showed you before now that I've deployed this all I need to do to start interacting with this within my application is to head over to read me and then I can go create a new instance I can call it YouTube example 0202 and I can create this new deployment where I can start interacting with it and here I have a bunch of different ways that I can interact with it I can either do it on the web here I can use a python script I can use typescript but most importantly if you're going to be doing what I'm showing you at the end of the video you're probably going to want to use this curl so this is going to give you all the information of where you need to connect to the endpoint uh the headers and also the data that it's expecting as well now I know this is a little bit more Technical and more code involved than the previous two but I want you to know that this is all very basic stuff and if you have an idea you can either come to myself you can get in touch with me I also do Consulting that I've watched recently so it's down there in the description and then the uh in the pin comment as well so if you have an idea and you want to get the spun up into a prototype you can have a chat with me I'll tell you how I do it and I can put you in touch with my developers to get it built out but essentially you can go from idea into a basic python script and then you take all this information here within this curl request then you can connect it to your front end as I'm going to show you at the end of this video now the fourth and

final way that you can create and deploy your first AI application is with berry AI now Berry is a little bit too convicts me to cover completely in this video so if you'd like to learn more about how I've been using Berry to spin up quick chatbot examples and how other people have been using to create entire applications and please let me know down below and I'll make an entire video for you breaking it down step by step very AI allows you to one click deploy your large language model apps it's a super fast and easy way to spin up in live website demo that you can essentially create a chat bot interface super easy and leverage all of their hard work they actually have an example here on their GitHub so if we head over to that and it's telling us to ask it if it can accept International payments from India so if I paste this in here it's been able to give me an answer to the question I just asked it the magic behind that little app just there is that all I've had to do is provide Barry with a link to where all that information is stored here you can see on screen is this docqa Pipeline and in order to initialize one of these all you need to do is provide your username your open AI key and then an input URL to the documentation or information that you want to train the model on what berry does is it takes all the data from the web page and then tokenizes that basically learns about it and then can allow you to query it in that chatbot setting this is a super easy way to spin up a quick prototype if you have a bunch of information say you want to make a customer support bot for your eCommerce store you give it a big list of all your customer service queries your frequently asked questions it's going to learn that instantly and then you can start playing around with it and asking it questions now aside from these very basic chatbot apps where you take a bit of information and teach it to the chatbot so you can interact with it Berry AI is extremely powerful because of its integration with Lang chain and GPT index I don't have time in this video to go over all the details on these but essentially Lang chain is a way for you to chain operations together and this is essentially how people are building the more complex large language models and AI apps at the moment land chain is the name might suggest allows you to connect a bunch of different operations together in a chain so while large language models like gpt3 and others are great on their own for different tasks typically you need to take and user input manipulate it do certain things and you can chain it all together with land chain I'm going to be making a land chain for dummies video very soon because I think it is it's very difficult to get your head around initially but once you understand it is essentially how all the impressive apps are being built right now and understanding it as an entrepreneur where some are looking to get into artificial intelligence app development is something you really need to

know and now finally we can get on to how we can take what we've just created with promptable every prompt steamship Etc and how can we actually put that onto our website so that we can start interacting it and potentially start monetizing it so to do this in the simplest way possible for you guys to understand we're going to be doing it on bubble. io now I'm logged in here and I'm going to create an app called example site for video I'm just going to skip all that and put in a hitter and a form so here we have our two elements now so on this side I'm going to be using that advice generator based on name and age so if you see on screen here I've got the YouTube example generate and all the information here so this is how I'm going to interact with this API input so what we need to do to interact with the application that I made in steamship is to provide it with a value for age and a value for name age rename all of these name label it's super important to name your input fields and your button so you know what you're looking at later on then we're going to get some text to display the output leave it like that we can Center it because it's a rich text editor we can Center the text great okay so what we're going to need to do now is head over to the plugins we're going to add a plugin which is the API connector and install that we're going to add another API we'll call this steamship and then we're going to copy this URL here which belongs can expand this now depending on your application your curve request here might be different but you need to find it you need to take note of all the different aspects of it the URL the headers and then you need to start putting this into the bubble back in I'm going to start by taking the URL here and I see it's a post request so I can see a post here and paste in the URL and I can change users to an action instead of data and now anything with a h tag is a header so I'm going to start copying these over add a header key and value foreign key and now we can see this D which is the data you want to copy that without the quotations and now we have the name and the value so double AI connector allows you to put in Dynamic content so here we can let it know we're going to insert something from our website we're planning to put something in there that is not going to be hard coded we can do it within these angle brackets so I'm going to put name here and then age here of course so what we can do now we have everything written up now we can initialize this call by providing it some example data so if I put my name in my agent now in this case steamship is trying to return me some text not Json so I need to make sure I've set my data type as text here then if I scroll down and I submit this re-initialize the call it's going to give me the output from this API call right next to it my advice for a 22 year old like you is to start thinking about your future now so we've verified that everything is working I'm able to connect to my steamship application everything is all ready to go and now I need to connect this API setup that I've done and connect it to my front end within bubble now what we need to do quickly here is just set up a state so Define another condition this text create a new state generated text so this is going to be a toggleable state that once our API call has returned and it's given us text we can then update the state so that it updates on the web page so I've got this text generated text when so that's the state is not empty then the element is going to be visible then I'm going to make sure that it's not visible and it's collapsed when hidden foreign if we go on to the button here double click it and go start edit workflow we can select this here which is steamship so when the button is clicked first step is to make that steamship API call and after that what we want to do is element actions set State and then search up submit button which we've named on the editor create a new custom state that is generated text State type text the element is the submit button custom state is generated text now the value all we need to do is leave it as the result of Step One steamship API call we need to click on this text box here and rename it to response text or anything that you like and then we insert the dynamic data by searching up response text generated text and there we go let's Center align this so it looks a bit nicer make it a little bit wider I'm sorry I just had to go back and fix a little area here I had the elements hit to the button you actually need to set it to the response text that we've named so response text generated text and the result is the result of step one which is the API calls so just to break down this workflow we're going to when the submit button is clicked it's going to take in the input and send it off with the API call it's going to update the state to contain the response text and then that's going to apply it to the response text element within the editor so it should all be working now if we go back and we preview this into my name and age and get advice my advice would be to make sure that you're taking care of yourself and your future there you have it guys we've gone through four different ways where you can create and deploy your very first AI application obviously very basic example but you get the idea bubblers are very good method for if you don't have much experience technically to spin up a quick prototype and you can easily Host this on a domain just buy a domain connect and then you have your first application up and ready to go that's all for today guys I hope you got something out of it I've given you four different ways that you can deploy and create your first AI application now I know it was very simple in the beginning and we got a bit complex with bubble but if there's anything that was unclear and you want me to go over it again in another video I can definitely do that for you guys I'm going to be doing my land Chain video but if you want to see more about Berry Ai and how you guys can create a quick chat bot for your business or for a different purpose let me know and I'll make a video for you guys I am here to serve you so please if you've enjoyed the video leave me a like it means a ton and if you like this kind of AI entrepreneurship focused content and these sort of Loco tutorials then please hit down below subscribe to the channel and hit the Bell to be notified when I release my next one that's all for the video thank you so much for watching guys and I'll see you in the next one